Revenue Is a Rumor Until It Settles
Payments taught me a lesson early that I have not been able to unlearn since. It is this: the number on the dashboard is not money. It is a rumor about money. Money is what is in the bank after everyone who can reach into the transaction has finished reaching.
Founders outside of payments rarely believe this until it hurts them, and it hurts a surprising number of them, because the tools have made the rumor look more and more like the fact.
The distance between booked and banked
Walk through what actually happens when a customer pays you. They click. A processor authorizes the amount, which means a bank has agreed the card is probably good. Your dashboard lights up. Your revenue chart moves. Somebody on the team posts a screenshot.
None of that is money yet. The funds settle days later, minus the processor's cut, minus the network's cut, minus whatever reserve your provider holds against you because you are new and they do not trust you. Then the customer has weeks, often months, to dispute the charge, and if they do, the money leaves your account before anyone asks your opinion, and a fee leaves with it. If the customer is on a subscription, the next charge can fail for a dozen reasons that have nothing to do with whether they still want the product. If you sell across borders, the number you booked and the number that arrives differ by whatever the currency did in between.
Booked revenue is the top of that funnel. Banked revenue is the bottom. In a healthy business the two are close. In a business that is about to have a very bad quarter, they have been drifting apart for months, and nobody noticed, because the dashboard only shows the top.
Where I have watched this go wrong
I have seen versions of the same failure in payments, in digital currency, and lately in AI companies that bill by usage, and the pattern is the same each time.
A company raises money on booked revenue and hires against it. The hires are real. The payroll is real. The revenue turns out to be a good deal smaller once refunds, disputes, failed renewals and reserves are through with it, and the gap between what was hired for and what arrived becomes the runway that was not there.
A company reports growth to its board that is actually growth in authorizations. Disputes lag by a couple of months, so the quarter that looked best is the one whose chargebacks land in the next one, and the board meeting where everyone celebrated is followed by the board meeting where everyone asks what happened.
A company treats a reserve as cash. Processors hold back a percentage of volume from young or risky merchants, sometimes for months. It is your money, in a sense. It is not your money in the only sense that matters when payroll runs.
And the one that is newest, and that I think will define the next few years: a usage-billed AI product books revenue every time a customer's agent calls its API, and the customer's finance team discovers, at the end of the month, that an agent ran in a loop for a week. The dispute arrives with a lawyer attached. The revenue was never real. The compute bill was.
Rules I now run every company by
Report banked, not booked, to yourself. Let the marketing team use the big number. The number you plan against, hire against and sleep against is the one that has cleared and survived the dispute window. If you do not know that number, you do not know your revenue, whatever the chart says.
Know your settlement lag to the day. How long between a customer paying and the money being yours to spend? Most founders guess "a couple of days" and the true answer, once reserves and disputes are included, is closer to a quarter. Your cash planning is only as good as that one number.
Track the funnel, not the total. Authorized, settled, disputed, refunded, failed, held. Six numbers, every week. The interesting information is never in the total. It is in which stage moved.
Treat the dispute rate as a product metric. Customers dispute charges they do not recognize, did not expect, or could not cancel. Every one of those is a design failure that showed up in the finance department. The payments team should be in the product review, and the product team should read the dispute reasons.
Never let the dashboard be the ledger. The dashboard is a view. The ledger is the truth, and the truth reconciles to the bank statement. The day those two disagree is the day you find out which one your company was actually running on.
Why this is harder now, not easier
Better tooling should have fixed this. It has made it worse. A modern billing stack produces a beautiful real-time revenue chart within an hour of setup, and the chart is so good that it stops being a view and becomes the number. Meanwhile the settlement cycle is as slow as it was twenty years ago, because it is bounded by banks and card networks and dispute law, none of which care about your dashboard's refresh rate.
So the gap between how fast the rumor arrives and how slowly the money does has never been wider, and the founders who get hurt are not the careless ones. They are the ones who trusted a very well-built instrument that was measuring the wrong thing.
Money is real when it settles. Everything before that is a story you are telling yourself about the future, and the future, in my experience, takes its cut.